How does the REPLACE_MAPPING dataset work for special characters?
In the 2022 version, we see that when we import a user with certain special characters those characters are not changed. Then if you update any attribute on the user, it seems to fire off a replace of special characters with the replace_mapping data...